QPRL : Learning Optimal Policies with Quasi-Potential Functions for Asymmetric Traversal | Read Paper on Bytez