Metraipolis › archive › dc_dbf5988c2e72

AccretionLink: On-Device Auditing of Exposure-Control Attacks on Attribute Inference — arXiv:2608.14735

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.14735
License (as archived)
CC0-1.0
Archived
2026-10-18T19:33:48.088Z
SHA-256 of the archived text
251f0569b00b08e072ce11e739890ece70011f339efa084e57fa7d18723bcd8a
Stable URL of this page
https://metraipolis.com/documento/dc_dbf5988c2e72

How to cite

AccretionLink: On-Device Auditing of Exposure-Control Attacks on Attribute Inference — arXiv:2608.14735 — arXiv (en), https://arxiv.org/abs/2608.14735 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_dbf5988c2e72 (Metraipolis archive dc_dbf5988c2e72, sha256 251f0569b00b08e072ce11e739890ece70011f339efa084e57fa7d18723bcd8a)
BibTeX
@misc{dcdbf5988c2e72,
  title = {AccretionLink: On-Device Auditing of Exposure-Control Attacks on Attribute Inference — arXiv:2608.14735},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.14735},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_dbf5988c2e72; sha256 251f0569b00b08e072ce11e739890ece70011f339efa084e57fa7d18723bcd8a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.14735. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Exposure control lets an adversary rank authentic public posts to strengthen private-attribute inference without altering content. AccretionLink defines confidentiality and integrity games for this attack, models bounded selection odds through partial identification, and constructs dependence-aware time-uniform e-processes. On 52 held-out synthetic profiles, odds-four selection reduced aggregate negative log likelihood at every horizon. At eight posts the advantage was 0.01595 nats (95% CI [0.00890, 0.02336]), three of four target effects survived Holm adjustment, and label-blind model-guided selection caused 6/109 high-confidence false reversals. On 142 PAN15 test profiles, exploratory selection produced a 0.01227-nat advantage but no reversal. A separate TF-IDF selector retained a 0.01470-nat advantage against the unchanged G5 target, while matched identity shuffling did not reproduce it. Pixel 10 encoded all 1,622 held-out posts once with a fallback-free Tensor G5 graph. A P-256 checkpoint authenticated the selected-replay, actual-model, native-report, and operation digests; local KeyInfo identified the signing key as StrongBox-backed.

Tema degli esploratori: fiducia digitale e crittografia (retro-catalogo arXiv: annunci del 2026-08-18 in cs.CR).
Categorie arXiv: Cryptography and Security (cs.CR) ; Machine Learning (stat.ML).
Pagina del lavoro: https://arxiv.org/abs/2608.14735 — 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 2026-10-18T19:01:59.100Z dalla pagina https://arxiv.org/catchup/cs.CR/2026-08-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).