An Efficiency-Preserving Transformation from Honest-Verifier Statistical Zero-Knowledge to Statistical Zero-Knowledge.

We present an unconditional transformation from any honest-verifier statistical zero-knowledge (HVSZK) protocol to standard SZK that preserves round complexity and efficiency of both the verifier and the prover. This improves over currently known transformations, which either rely on some computational assumptions or introduce significant computational overhead. Our main conceptual contribution is the introduction of instance-dependent SZK proofs for NP, which serve as a building block in our transformation. Instance-dependent SZK for NP can be constructed unconditionally based on instance-dependent commitment schemes of Ong and Vadhan (TCC'08). As an additional contribution, we give a simple constant-round SZK protocol for Statistical-Difference resembling the textbook HVSZK proof of Sahai and Vadhan (J.ACM'03). This yields a conceptually simple constant-round protocol for all of SZK.