SQIL: Imitation Learning via Reinforcement Learning with Sparse Rewards - Citation Graph | Papersgraph