UC Berkeley

Zero-knowledge proof is a cryptographic protocol enabling provers to convince verifiers of a statement's validity without disclosing any additional information beyond its truthfulness. It can be used to address security and privacy challenges in diverse fields, such as artificial intelligence, data analytics, and blockchain.

In this dissertation, we delve into the zero-knowledge proof, examining both its theoretical foundations and real-world applications. In theory, we introduce Libra, Virgo and Virgo++, a series of pioneering zero-knowledge protocols that boast optimal prover time, rapid verifier time, and succinct proof size while maintaining exceptional efficiency in practical settings. In practice, we explore the groundbreaking implementation of zero-knowledge proof in creating secure, trustless, and permissionless cross-chain bridges for blockchain networks. Furthermore, we also investigate the application of zero-knowledge proof in the realm of machine learning, showcasing its potential to ensure the integrity of machine learning models, as demonstrated with the decision tree model. The applied zero-knowledge proof protocols presented in this dissertation offer robust security assurances coupled with pragmatic efficiency.