Pareto-optimal quantum kernel selection for unsupervised anomaly detection on real malware beaconing data

Abstract

Quantum kernel methods are leading candidates for a practical quantum advantage in machine learning, but assessing that potential requires two quantities usually reported separately: how well a kernel performs on the task, and how far its geometry departs from the classical kernels available for the same problem. We introduce a fully unsupervised, multi-objective protocol that optimises simultaneously the normalised pseudo discrepancy (NPD), a label-free proxy for anomaly detection quality, and the geometric difference (GD) to a tuned classical reference kernel, selecting models from the resulting Pareto front. We apply it to malware beaconing detection in real network traffic, using a one-class support vector machine with fidelity and projected quantum kernels over four data encodings, on simulators and on IQM's 20-qubit Garnet processor. NPD-guided selection alone finds a fidelity kernel that beats the tuned classical baseline, but with a geometric difference too small to certify the gain as quantum. Projected kernels reach far larger geometric differences; the Pareto-selected one only marginally exceeds the baseline (AUC $0.782$ versus $0.765$, g_{C→ Q}≈ 89>√N relative to that reference kernel), still below the NPD-selected fidelity kernel ($0.840$).

Bliss of dimensionality: How an unsupervised criterion identifies optimal low-resolution representations of high-dimensional datasets
Margherita Mele, Daniel Campos, Raffaello Potestio
Physical Review Research26 Oct 2026Peer-reviewedCC BY 4.0Atomic, molecular & optical physicsAPS
Landscape-Dependent Performance of Photonic Quantum Solvers in QUBO Feature Selection for Financial Risk Detection
Nirvik Sahoo, Paul Robert Griffin
arXiv2 Oct 2026PreprintCC BY 4.0Full textQuantum physics & quantum informationarXiv
Exoplanet Detection Using Adaptive Quantum-Optimal Measurement
Hyunsoo Choi, Hyoung Won Baac, Zubin Jacob +1
arXiv22 Sep 2026PreprintarXiv licenceFull textInstrumentation & methods Planetary science & exoplanetsarXiv
Coarse-Graining Hidden Representations: Unsupervised Neuron Selection via Mapping Entropy
Margherita Mele, Andrea Castagna, Roberto Menichetti +2
arXiv4 Sep 2026PreprintCC BY 4.0Full textCondensed matter Statistical & nonlinear physicsarXiv
Likelihood-Based Unsupervised Anomaly Detection in CMS Dijet Events
Bhavishya Chebrolu, Hitesh Rasineni, Prajwal Aaryan Immadi
arXiv6 Sep 2026PreprintarXiv licenceFull textHigh-energy & particle physicsarXiv

Cite

BibTeX RIS

BibTeX

@misc{micah2026paretooptimal,
  title = {{Pareto-optimal quantum kernel selection for unsupervised anomaly detection on real malware beaconing data}},
  author = {Boaz Micah and Nadia Milazzo and Maissa Beji and Borja Aizpurua and Llorenç Espinosa-Portalés and Esteban Payares and Ghada Ben Slama and Luc Andrea and Michel Kurek and Thomas Cope and Olivier Salomon},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.09717},
  eprint = {2610.09717},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.09717},
}

RIS

TY  - UNPB
TI  - Pareto-optimal quantum kernel selection for unsupervised anomaly detection on real malware beaconing data
AU  - Boaz Micah
AU  - Nadia Milazzo
AU  - Maissa Beji
AU  - Borja Aizpurua
AU  - Llorenç Espinosa-Portalés
AU  - Esteban Payares
AU  - Ghada Ben Slama
AU  - Luc Andrea
AU  - Michel Kurek
AU  - Thomas Cope
AU  - Olivier Salomon
PY  - 2026
DA  - 2026/10/07
PB  - arXiv
DO  - 10.48550/arxiv.2610.09717
UR  - https://arxiv.org/abs/2610.09717
AB  - Quantum kernel methods are leading candidates for a practical quantum advantage in machine learning, but assessing that potential requires two quantities usually reported separately: how well a kernel performs on the task, and how far its geometry departs from the classical kernels available for the same problem. We introduce a fully unsupervised, multi-objective protocol that optimises simultaneously the normalised pseudo discrepancy (NPD), a label-free proxy for anomaly detection quality, and the geometric difference (GD) to a tuned classical reference kernel, selecting models from the resulting Pareto front. We apply it to malware beaconing detection in real network traffic, using a one-class support vector machine with fidelity and projected quantum kernels over four data encodings, on simulators and on IQM's 20-qubit Garnet processor. NPD-guided selection alone finds a fidelity kernel that beats the tuned classical baseline, but with a geometric difference too small to certify the gain as quantum. Projected kernels reach far larger geometric differences; the Pareto-selected one only marginally exceeds the baseline (AUC $0.782$ versus $0.765$, g_C→ Q≈ 89>√N relative to that reference kernel), still below the NPD-selected fidelity kernel ($0.840$).
ER  -

Abstract

Related

Cite