PixL2R: Guiding Reinforcement Learning Using Natural Language by Mapping Pixels to Rewards.

Prasoon Goyal, Scott Niekum, Raymond J. Mooney

Browse the full CoRL paper archive.