Curator's Take
AI Commentary
This article demonstrates that even a modest four‑qubit ZZ feature‑map can retain most of its intended geometric structure when run on real IBM hardware, with gate‑twirling providing the clearest preservation across several diagnostic metrics. By quantifying how much of the statevector‑referenced kernel survives under baseline, dynamical decoupling and twirling, the work offers a concrete benchmark for assessing quantum‑kernel fidelity in noisy intermediate‑scale devices—a key step toward reliable QML pipelines. The findings also highlight that higher geometric fidelity does not automatically translate into better alignment with target labels, underscoring the nuanced trade‑offs that must be managed when deploying quantum kernels on today’s hardware.
— Mark Eatherly
Summary
Quantum-kernel methods encode a dataset's geometry in a Gram matrix, so learning claims on hardware kernels assume the intended geometry survives execution. We measure that survival for one frozen four-qubit ZZ feature-map kernel on $N=24$ real indoor air-quality windows, reconstructed on ibm_fez (1024 shots per circuit) under baseline, dynamical decoupling alone, and gate twirling alone, each a single non-interleaved job. Every configuration returned a complete, finite, positive-semidefinite Gram matrix and preserved the centered statevector geometry to a substantial but incomplete descriptive degree (full-matrix centered kernel alignment, CKA, 0.933-0.989). Gate twirling was most faithful on every reported geometry axis, with the only jackknife-resolved improvement over baseline (persisted Spearman, mean absolute error, and full-matrix CKA diagnostics); dynamical decoupling alone was not separated from baseline at the frozen-window scale. Residual hardware distortion, not finite sampling, dominates the discrepancy. Yet fidelity and label alignment were reversed: the most faithful configuration had the lowest centered kernel-target alignment, which sits at or below label-permutation references for statevector and hardware alike. We read the small hardware uplift as a normalization property of the non-affine distortion, not captured signal. These are descriptive results for single jobs on one backend, not causal mitigation-efficacy estimates; no quantum-advantage, hardware-classifier-superiority, or forecasting claim is made. Implementation fidelity and task relevance are distinct axes; hardware quantum machine-learning studies should report both.