Privacy-Preserving and Verifiable Multi-Armed Bandit Decision Process via zk-SNARKs

This study introduces zkUCB, an innovative algorithm that integrates the Zero-Knowledge Succinct Non-Interactive Argument of Knowledge (zk-SNARKs) with the Upper Confidence Bound (UCB) algorithm to enable privacy-preserving and verifiable decision-making in Multi-Armed Bandit (MAB) problems.