Self-Supervised Online Reward Shaping in Sparse-Reward Environments | Read Paper on Bytez