Metraipolis › archive › dc_1e49e694fcb4
Adaptive Model Inversion Attacks Generalize a Privacy-Robustness Tradeoff — arXiv:2610.07677
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.07677
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T11:35:26.271Z
- SHA-256 of the archived text
a03a78578cdcc43914edbab92df94a70342d823800da4c7128580be72cbdfccc- Stable URL of this page
https://metraipolis.com/documento/dc_1e49e694fcb4
How to cite
Adaptive Model Inversion Attacks Generalize a Privacy-Robustness Tradeoff — arXiv:2610.07677 — arXiv (en), https://arxiv.org/abs/2610.07677 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1e49e694fcb4 (Metraipolis archive dc_1e49e694fcb4, sha256 a03a78578cdcc43914edbab92df94a70342d823800da4c7128580be72cbdfccc)
BibTeX
@misc{dc1e49e694fcb4,
title = {Adaptive Model Inversion Attacks Generalize a Privacy-Robustness Tradeoff — arXiv:2610.07677},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.07677},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1e49e694fcb4; sha256 a03a78578cdcc43914edbab92df94a70342d823800da4c7128580be72cbdfccc},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.07677. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
In this paper, we show that standard evaluations of high-resolution Model Inversion Attacks (MIAs) significantly underestimate training-data privacy leakage. State-of-the-art privacy defenses, standard training techniques such as MixUp and Adversarial Training, and undefended models all leak training images at rates 1.16 to 6.59 times higher on FaceScrub under simple adaptive changes to the attack, with the largest increases among defenses reporting the strongest privacy. We further show that measured leakage depends on the feature basis of the external classifier used to evaluate reconstructions: for the same reconstructed images, an adversarially trained Inception evaluator identifies the targeted identity at different rates than the standard Inception evaluator. Our results suggest that standard MIA evaluation can mistake optimization and measurement failures for privacy. These underestimated leakage rates also concealed a broader relationship between privacy and adversarial robustness. Once we adapt the attack and vary the evaluator, reconstruction leakage closely tracks adversarial robustness across recent defenses and standard training regimes, suggesting that robustness provides an attack-agnostic proxy for reconstruction vulnerability that applies far more broadly than previously theorized. This raises an open question: can a practical defense reduce training-data reconstruction without paying a corresponding cost in adversarial robustness? Tema degli esploratori: fiducia digitale e crittografia (elenco arXiv cs.CR, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG) ; Cryptography and Security (cs.CR). Pagina del lavoro: https://arxiv.org/abs/2610.07677 — 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 2026-10-07T11:18:39.723Z dalla pagina https://arxiv.org/list/cs.CR/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).