Security Evaluation of Quantum Circuit Split Compilation under an Oracle-Guided Attack

Hongyu Zhang and Yuntao Liu · 2025

Read Security Evaluation of Quantum Circuit Split Compilation under an Oracle-Guided Attack by Hongyu Zhang and Yuntao Liu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-04842.txt.

Introduction

Hongyu Zhang and his co-author test whether split compilation truly protects quantum circuit designs. Quantum circuits are valuable intellectual property, and split compilation hides their wiring. The circuit is cut into two parts, each visible, while the links between them stay secret.

Their oracle-guided attack tries to recover those hidden connections systematically. Candidate links are proposed, checked against an oracle, and refined step by step. A full security evaluation framework wraps the attack for repeatable testing. The work warns compiler designers that splitting alone may leak more than assumed, so obfuscation claims need adversarial proof.

Catalog record

CREATORHongyu Zhang and Yuntao Liu
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE27 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-04842

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Security Evaluation of Quantum Circuit Split Compilation under an Oracle-Guided Attack Hongyu Zhang and Yuntao Liu Abstract Quantum circuits are the fundamental representation of quantum algorithms and constitute valuable intellectual property (IP). Multiple quantum circuit obfuscation (QCO) techniques have been proposed in prior research to protect quantum circuit IP against malicious compilers. However, there has not been a thorough security evaluation of these schemes. In this work, we investigate the resilience of split compilation against an oracle-guided attack. Split compilation is one of the most studied QCO techniques, where the circuit to be compiled is split into two disjoint partitions. Each split circuit is known to the compiler, but the interconnections between them are hidden. We propose an oracle-guided security evaluation framework in which candidate connections are systematically tested against input-output observations, with iteratively pruned inconsistent mappings. This hierarchical matching process exploits the reversibility of quantum gates and reduces the search space compared to brute-force enumeration. Experimental evaluation in the RevLib benchmark suite shows that only a small number of I/O pairs are sufficient to recover the correct inter-split connections and reconstruct the entire circuits. Our study marks the first thorough security evaluations in quantum IP protection and highlights the necessity of such evaluations in the development of new protection schemes. I Introduction The advancement of quantum computers ushers in a new era of computational power and methodologies. Quantum algorithms are implemented through single operations, called quantum gates, while gate sequences form quantum circuits, the hardware-executable representation of quantum programs. Cloud platforms such as IBM Quantum, Amazon Braket, and Microsoft Azure [ 5 , 8 , 18 ] provide extensive access to quantum hardware, while compilers such asskit [ 20 ] map high-level circuit descriptions to device-specific instructions. Despite these advances, quantum circuit design remains resource-intensive and requires substantial intellectual capital. Consequently, quantum circuit designs are increasingly recognized as valuable intellectual property (IP) [ 1 ] , analogous to integrated circuits (ICs) in classical computing. Recent research has explored quantum-specific defense mechanisms—including split compilation [ 23 , 28 ] , insertion of randomized reversible subcircuits [ 7 , 26 ] , quantum logic locking [ 27 , 13 ] , and phase-level obfuscation techniques [ 22 ] to protect against threats to designs posed by untrusted compilers and cloud environments. Split compilation [ 23 , 28 ] , inserting random reversible circuits [ 7 , 26 ] , quantum logic locking [ 27 , 13 ] , and phase obfuscation [ 22 ] to protect against untrusted compilers capable of IP theft or Trojan insertion [ 6 , 11 ] . These efforts assume that hiding or fragmenting circuit structure is sufficient to prevent adversaries from recovering original designs. However, these works did not consider any sophisticated adversaries beyond brute-force attacks. However, a significant gap remains: the problem of reconstructing circuits when their partitioned forms are exposed, but interconnections remain hidden. For example, the work on split compilation [ 23 , 28 ] assumed that adversaries needed to exhaustively try all the permutations of qubit mappings between the two split circuits. This oversimplified assumption overlooked the structural and functional information that exists in the split circuits that can be leveraged to reduce the number of I/O queries needed to reconstruct the qubit mapping in the presence of a black-box oracle. We propose a reverse engineering framework for quantum circuits protected by partitioned compilation. Attackers who know both partitioned circuits but lack knowledge of the qubit mapping can determine the circuit correspondence by selecting inputs and measuring outputs. We employ a hierarchical reconstruction algorithm to process split-compiled circuits. The process first stitches single-wire traces, then progressively expands to dual-layer, triple-layer, and ultimately complete boundary replacement, that is, the entire circuit, until parsed. At the reconstruction level k ( k=1,2,\ldots ), we generate candidate k wire interconnect schemes while maintaining single-wire consistency; The surviving set \mathcal{H}_{k} undergoes iterative validation through queries to a finite budget input-output prediction oracle, eliminating any hypotheses that fail input-output consistency checks. Using gate reversibility and systematic block elimination mechanisms, this algorithm significantly reduces search space compared to brute-force methods that exhaustively explore all boundary permutations. The main contributions of this paper are as follows. • Problem formulation and query efficiency. We transform the reverse engineering problem of split-compiled circuits into connection restoration rather than an exhaustive permutation search. This reduces the number of required oracle queries under ideal conditions. • Hierarchical reconstruction. We employ a layered approach, starting with single-wire consistency within each partitioned unit and progressively expanding to pairwise and multi-wire cross-partition hypotheses until the entire circuit is resolved. At each level, the candidate interconnections are validated against the observed input-output ( x,y ) behavior of the original circuit. • Security evaluation. To our knowledge, this is the first systematic security evaluation of quantum partitioned (split) compilation as an intellectual property protection technique, conducted on standard benchmarks under controlled query budgets. Our proposed framework reconstructs the implicit interconnections between partitioned layers. This process first connects single paths within each partitioned layer, then progressively expands to establish paired and multipath connections between two partitioned layers, ultimately converging into a complete circuit. At each stage, candidate connections are validated using all input-output data across the circuit while eliminating inconsistent mapping relationships. Using the reversibility of quantum gates and employing a systematic block pruning strategy, this method significantly reduces the search space compared to brute-force enumeration. Experiments on the RevLib benchmark demonstrate that only a small number of input-output pairs are required to recover the correct inter-segment connections. This approach substantially improves the efficiency of circuit reconstruction. This research demonstrates the feasibility of reconstructing hidden interconnections between circuit segments and highlights the necessity of strengthening quantum intellectual property protection defense mechanisms in practical deployments. II Background & Related Work II-A Quantum Computing Fundamentals Quantum computing is one of the top five priority areas identified by the National Science Foundation for future investment and has already demonstrated clear advantages over classical computing in specific domains, marking significant milestones in computational science. One of the most notable achievements is Google’s 2024 demonstration of quantum supremacy using its 105-qubit Willow processor [ 16 ] . Willow completed a random circuit sampling task in under five minutes, while the same task would take classical supercomputers about 10 septillion years. This shows that quantum devices have shown the ability to outperform classical systems in computational tasks, including practical advantages in unstructured search, period finding, optimization, and machine learning. Many quantum algorithms promise asymptotic speedups over their best-known classical counterparts, ranging from polynomial to exponential improvements. For example, the Quantum Approximate Optimization Algorithm (QAOA) targets combinatorial optimization with potential polynomial advantages [ 10 , 15 ] , while the Variational Quantum Eigensolver (VQE) has shown promise in quantum chemistry applications [ 17 , 12 ] . At the other extreme, Shor’s algorithm achieves an exponential speedup for factoring and discrete logarithms using the Quantum Fourier Transform (QFT) [ 24 , 19 ] , and Grover’s algorithm provides a quadratic advantage for unstructured search [ 9 ] . Quantum computing relies on quantum bits, a.k.a. qubits, which differ from classical bits in that they can simultaneously exist in a superposition state of the ground states \ket{0}

Read the complete text