Metraipolis › archive › dc_a79a1bb71d4c

When Normalization Selects the Sign: Auditing Robustness Ablations in Quantum Attention — arXiv:2610.02641

Source
arXiv
Original
https://arxiv.org/abs/2610.02641
License (as archived)
CC0-1.0
Archived
2026-10-05T23:06:29.316Z
SHA-256 of the archived text
732ecd2b1200ada857d9577ff6029611fec919e996a467673b39eaa7f6e0b640
Stable URL of this page
https://metraipolis.com/documento/dc_a79a1bb71d4c

How to cite

When Normalization Selects the Sign: Auditing Robustness Ablations in Quantum Attention — arXiv:2610.02641 — arXiv, https://arxiv.org/abs/2610.02641 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a79a1bb71d4c (Metraipolis archive dc_a79a1bb71d4c, sha256 732ecd2b1200ada857d9577ff6029611fec919e996a467673b39eaa7f6e0b640)
BibTeX
@misc{dca79a1bb71d4c,
  title = {When Normalization Selects the Sign: Auditing Robustness Ablations in Quantum Attention — arXiv:2610.02641},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.02641},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a79a1bb71d4c; sha256 732ecd2b1200ada857d9577ff6029611fec919e996a467673b39eaa7f6e0b640},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.02641. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Removing an input-scaling module changes both a classifier and the perturbations reaching its encoder. A robustness difference can therefore reflect the comparison rule as well as the module. We demonstrate this problem in a four-qubit quantum-attention detector on generated power-grid trajectories. A learned scaling module appears beneficial at a fixed physical attack budget, but matching an upper bound on perturbations at the encoder reverses the ordering. Neither comparison alone establishes a robustness benefit caused by the module. The initial test also perturbs clean examples into attacked examples while retaining their original labels; tests restricted to already attacked examples do not establish a benefit. Replacing a trained model's input scales disrupts detection. Retraining its linear classification layer restores the detection rate, but changes individual predictions, leaving the comparison descriptive rather than causal. Two further design checks explain why the input quantum Fisher information regularizer cannot train this model's query parameters, and why removing confidence bounds does not establish a larger certified radius. The evidence is limited to ten seeds, exact simulation, synthetic data, and a restricted set of attacks; classical baselines achieve better clean prediction. The practical lesson is to specify which perturbation budget is fixed, check that attacks preserve labels and interventions preserve predictions, and distinguish exploratory controls from confirmatory evidence.

Tema degli esploratori: fiducia digitale e crittografia (elenco arXiv cs.CR, nuovi lavori del giorno).
Categorie arXiv: Quantum Physics (quant-ph) ; Cryptography and Security (cs.CR); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.02641 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-05T23:01:08.555Z dalla pagina https://arxiv.org/list/cs.CR/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).