An Investigation into Pre-Training Object-Centric Representations for Reinforcement Learning.

Jaesik Yoon, Yi-Fu Wu, Heechul Bae, Sungjin Ahn

Browse the full ICML paper archive.