Bayesian Optimization from Human Feedback: Near-Optimal Regret Bounds | Read Paper on Bytez