On adaptively secure protocols

Abstract

Adaptive security captures the capability of an adversary to adaptively affect a system during the course of its computation based on partial information gathered. In this work, we explore the theoretical complexity of achieving adaptive security in two settings: 1. Adaptive UC-Secure Computation: We provide a round-efficient compiler that transforms any stand-alone semi-honest adaptively secure multiparty computation to adaptive UC-security. Recently, Dana et. al (Asiacrypt 2013) showed how to acheive adaptive UC-security in any trusted setup under minimal assumptions. They achieve this by constructing an O(n)-round adaptively secure concurrent non-malleable commitment scheme. The main contribution of our work shows how to achieve the same in O(1)-rounds. 2. Zero-Knowledge with Adaptive Inputs: Lin and Pass in (TCC 2011) gave first constructions of concurrent non-malleable zeroknow- ledge proofs secure w.r.t. adaptively chosen inputs in the plain model in a restricted setting, namely, where the adversary can only ask for proofs of true (adaptively-chosen) statements. We extend their definition to the fully-adaptive setting and show how to construct a protocol that satisfies this definition. As an independent contribution we provide a simple and direct compilation of any semihonest secure protocol to a fully concurrently secure protocol under polynomial-time assumptions in the Angel-Based UC-Security.