Metraipolis › archive › dc_abab9fc5e346
On the Relationship between Model Quantization and Model Inversion Attacks — arXiv:2610.00382
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.00382
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-06T17:48:35.380Z
- SHA-256 of the archived text
b93545ac460200c3835cea47a794f0519afe03682d56cfdca84910ad0f7c4ba0- Stable URL of this page
https://metraipolis.com/documento/dc_abab9fc5e346
How to cite
On the Relationship between Model Quantization and Model Inversion Attacks — arXiv:2610.00382 — arXiv (en), https://arxiv.org/abs/2610.00382 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_abab9fc5e346 (Metraipolis archive dc_abab9fc5e346, sha256 b93545ac460200c3835cea47a794f0519afe03682d56cfdca84910ad0f7c4ba0)
BibTeX
@misc{dcabab9fc5e346,
title = {On the Relationship between Model Quantization and Model Inversion Attacks — arXiv:2610.00382},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.00382},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_abab9fc5e346; sha256 b93545ac460200c3835cea47a794f0519afe03682d56cfdca84910ad0f7c4ba0},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.00382. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Model quantization reduces the numerical precision of neural network weights and activations to lower storage and computational costs. Model inversion attacks recover or reconstruct sensitive training data or inference inputs from model outputs or intermediate features, so quantization may also alter their effectiveness. However, two questions remain unresolved: How does model quantization affect model inversion? How do data characteristics influence this relationship? To address the first, we bound quantization-induced changes in mutual information between inputs and a categorical variable defined by prediction probabilities, distinguishing informational effects from attack optimization obstacles. To address the second, we identify data-dependent changes in feature distributions and inversion outcomes, with pronounced quantization sensitivity differences at 4 bits. These insights guide a privacy-aware post-training quantization method that improves inversion resistance while recovering utility. It uses a Fisher-type task-sensitivity proxy for budget-aware bit allocation, calibrates activation ranges, and jointly optimizes weight and activation scales and weight-rounding decisions with task-recovery and geometry-retention objectives and scale and rounding regularization. Experiments cover multiple metrics, neural network architectures, and face, palmprint, and iris recognition tasks. On ResNet-50, Palm at 4 bits reduces RL-MIA's strict success from 54% to 26%, while accuracy decreases from 99.01% to 96.55% relative to FP32. Our method also supports output-level defenses: adding Stealthy Shield Defense (SSD, epsilon = 0.1) to Iris at 4.5 bits reduces BREP-MI's strict success from 63.33% to 37.33%, while accuracy decreases from 92.8% to 87.6% relative to quantization alone. Tema degli esploratori: matematica e informatica teorica (retro-catalogo arXiv: annunci del 2026-10-02 in math.IT). Categorie arXiv: Cryptography and Security (cs.CR) ; Information Theory (cs.IT); Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2610.00382 — 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 2026-10-06T16:13:30.447Z dalla pagina https://arxiv.org/catchup/math.IT/2026-10-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).