Efficient Zero-Knowledge Proofs and their Applications

A zero-knowledge proof is a fundamental cryptographic primitive that enables the verification of statements without revealing unnecessary information. Zero-knowledge proofs are a key component of many cryptographic protocols and, often, one of their main efficiency bottlenecks. In recent years there have been great advances in improving the efficiency of zero-knowledge proofs, bring them closer to wide deployability. In this thesis we make another step towards the construction of computationally-efficient zero-knowledge proofs. Specifically, we construct efficient zero-knowledge proofs for the satisfiability of arithmetic circuits for which the computational cost of the prover is only a constant factor more expensive than direct evaluation of the circuit. We also construct efficient zero-knowledge proofs to check the correct execution of (Tiny)RAM programs. In this case the computational cost for the prover is a superconstant factor larger than executing the program directly. Our proofs also support efficient verification and small proof sizes. For security, they rely on symmetric primitives and could potentially withstand attacks from quantum computers. On a different research direction, we look at group signatures, a fundamental primitive which relies on zero-knowledge proofs. A group signature enables users to sign anonymously on behalf of a group of users. In case of dispute a Manager can identify the author of a signature and potentially banish the user from the group. In this thesis we address the fundamental question of defining the security of fully dynamic group signatures, for which the users can join and leave at any time. Differently from other restricted settings, this case has been largely overlooked in the past. Our security model is general, does not implicitly assume existing design paradigms and captures the security of existing models for more restricted settings.