Implicit zero-knowledge arguments and applications to the malicious setting

Abstract

We introduce implicit zero-knowledge arguments (iZK) and simulation-sound variants thereof (SSiZK); these are lightweight alternatives to zero-knowledge arguments for enforcing semi-honest behavior. Our main technical contribution is a construction of efficient two-flow iZK and SSiZK protocols for a large class of languages under the (plain) DDH assumption in cyclic groups in the common reference string model. As an application of iZK, we improve upon the round-efficiency of existing protocols for securely computing inner product under the DDH assumption. This new protocol in turn provides privacy-preserving biometric authentication with lower latency.