On Adaptively Secure Protocols

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 zero-know-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 semi-honest secure protocol to a fully concurrently secure protocol under polynomial-time assumptions in the Angel-Based UC-Security.