Abstract
Quantum physical unclonable functions (QPUFs) are emerging as critical hardware solutions for secure cryptographic functionalities, including key exchange, message authentication, and entity identification. Recent theoretical work establishes that provable security against quantum polynomial-time adversaries requires QPUFs to be unitaries sampled uniformly from the Haar measure. However, implementing such unitaries requires exponential resources, creating a significant feasibility gap. In this work, we propose an efficient construction using unitary $t$-designs, termed QPUF$_t$. We modify existing security definitions to accommodate efficient constructions and demonstrate that QPUF$_t$ maintains provable security guarantees against bounded quantum polynomial adversaries with $t$-query access. This represents the first application of unitary $t$-designs for arbitrary $t$. Furthermore, we investigate noise-resilience against unitary noise, concluding that while some resilience is achievable as system size increases, practical implementation will likely require error mitigation or correction.
1. Introduction
Physical Unclonable Functions (PUFs) serve as "digital fingerprints" for hardware, leveraging microscopic physical variations from the manufacturing process to generate unique, unclonable responses. In the quantum realm, QPUFs extend these capabilities, offering theoretically superior security for quantum communication and authentication tasks.
The primary challenge in QPUF design is the requirement for the device to behave like a unitary matrix sampled uniformly from the Haar measure—essentially a truly random rotation in a high-dimensional Hilbert space. While this provides maximum security, implementing a Haar-random unitary on $n$ qubits requires a circuit depth that scales exponentially with $n$, making it physically impossible for practical systems.
2. Proposed Construction: QPUF$_t$
To bridge the gap between theoretical security and physical realizability, we propose using unitary $t$-designs. A $t$-design is a distribution of unitaries that mimics the statistical properties (moments) of the Haar measure up to the $t$-th order. If an adversary is limited to $t$ queries, they cannot distinguish a $t$-design from a truly Haar-random unitary.
Our construction utilizes a Random Quantum Circuit (RQC) model. We employ a 1D chain of qubits where nearest-neighbor gates are applied in a structured pattern. These gates are selected from an approximately universal set, allowing the circuit to explore the state space efficiently. Unlike Haar-random unitaries, these QPUF$_t$ devices can be implemented with a circuit depth that scales polynomially with the number of qubits and the security parameter $t$.
3. Redefining Security for Practicality
Because $t$-designs are efficiently constructible, we must update standard security definitions. We introduce the concept of "Practical Unknownness." This definition accounts for the fact that while a polynomial-time adversary could theoretically simulate a known efficient circuit, they cannot distinguish a specific, secret QPUF instance from a random one when given only limited black-box interaction. We prove that QPUF$_t$ retains "selective unforgeability," meaning an adversary cannot predict the response to a new challenge state even after observing multiple challenge-response pairs.
4. Numerical Validation and Results
We validated our construction through numerical simulations, measuring the "distinguishability" (via the diamond norm) between different QPUF instances. A high distinguishability is essential to ensure that two different devices do not produce the same response.
Simulations were conducted for systems of 4, 6, and 8 qubits with varying circuit depths. Even with shallow circuits (4 blocks of gates), the distinguishability was remarkably high:
- 4 Qubits: Distinguishability $\ge 1.99$
- 8 Qubits: Distinguishability $\ge 1.99999$
These results confirm that even relatively low-depth RQCs can effectively approximate the maximal distinguishability of Haar-random unitaries.
5. Noise Resilience and Limitations
We analyzed the impact of unitary noise—imperfections in the quantum gates themselves. Our analysis shows that while noise shifts the distribution, the device remains a $t$-design, though its closeness to the ideal Haar measure degrades.
One limitation of the current study is the focus on unitary noise rather than more complex decoherence (stochastic noise). As system sizes scale, the error rates per qubit must decrease to maintain security properties. Consequently, we conclude that integrating quantum error correction or mitigation will be necessary for large-scale, real-world QPUF deployment.
6. Conclusion
By replacing ideal Haar-random unitaries with unitary $t$-designs, we provide a path toward practical, provably secure quantum hardware. QPUF$_t$ offers a scalable solution for quantum authentication and cryptography, marking a significant milestone in the transition from theoretical quantum primitives to realizable security hardware.
Summary updated