Fetching the paper…

Zeroth-Order Optimization Meets Human Feedback: Provable Learning via Ranking Oracles · Around