🤖 AI Summary
This work addresses the fundamental problem of unconditionally verifying quantumness in small-scale devices—i.e., enabling a classical verifier to efficiently certify that a prover possesses computational capabilities beyond any classical machine, without relying on any assumptions from quantum mechanics. We present the first Proof of Quantumness (PoQ) protocol with unconditional soundness, requiring only an upper bound on the prover’s memory. Our approach combines Raz’s matrix inversion space lower bound with the bounded-storage model, leveraging modulo-2 arithmetic and a minimal quantum gate set containing exactly one non-Clifford gate. Key contributions are: (1) the first construction achieving a quadratic-to-exponential separation in memory requirements between honest provers and arbitrary adversaries; and (2) two polynomial-time verifiable protocols—a hardware-friendly variant using a single non-Clifford gate, and a stronger variant guaranteeing exponential memory separation under adversarial conditions.
📝 Abstract
A proof of quantumness (PoQ) allows a classical verifier to efficiently test if a quantum machine is performing a computation that is infeasible for any classical machine. In this work, we propose a new approach for constructing PoQ protocols where soundness holds unconditionally assuming a bound on the memory of the prover, but otherwise no restrictions on its runtime. In this model, we propose two protocols: 1. A simple protocol with a quadratic gap between the memory required by the honest parties and the memory bound of the adversary. The soundness of this protocol relies on Raz's (classical) memory lower bound for matrix inversion (Raz, FOCS 2016). 2. A protocol that achieves an exponential gap, building on techniques from the literature on the bounded storage model (Dodis et al., Eurocrypt 2023). Both protocols are also efficiently verifiable. Despite having worse asymptotics, our first protocol is conceptually simple and relies only on arithmetic modulo 2, which can be implemented with one-qubit Hadamard and CNOT gates, plus a single one-qubit non-Clifford gate.