Reproduce “Post-Cut Metadata Inference Attacks on Quantum Circuit Cutting Pipelines” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Samuel Punch, Krishnendu Guha, Utz Roedig
Researchers from University College Cork demonstrated a metadata inference attack on quantum circuit cutting, allowing a semi-honest cloud provider to deduce structural properties of client workloads from routinely logged classical fragment metadata. The attack achieved high accuracy, up to 0.960 for algorithm family and Hamiltonian k-locality inference, even when controlling for fragment size, and was validated on production quantum hardware.
Key insights
- The "routing tax," reflecting topology-dependent compilation overhead (depth inflation, 2Q gate count increase, compiled width), serves as the core physical mechanism enabling metadata leakage.
- Classical control-plane metadata, such as compiled fragment width and depth, offers a high-resolution and deterministic side channel, in contrast to QPU execution timing which often appears invariant due to dominant control-plane latencies.
- The information leakage is primarily "structure-dominated," revealing genuine signatures of the original workload's topology and partition strategy, even when coarse fragment size differences are equalized.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.