Regret-Free Reinforcement Learning for Temporal Logic Specifications | Read Paper on Bytez