Metraipolis › archive › dc_e73f8083074a

Denoising-Aware Inversion: Revealing Privacy Risks in Noise-Protected Text Embeddings — arXiv:2608.18610

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.18610
License (as archived)
CC0-1.0
Archived
2026-10-18T13:45:14.199Z
SHA-256 of the archived text
d1afe15a9c96f8080d5c4f4bf8d672b25dda8e76213afe0c5288805d24a2b612
Stable URL of this page
https://metraipolis.com/documento/dc_e73f8083074a

How to cite

Denoising-Aware Inversion: Revealing Privacy Risks in Noise-Protected Text Embeddings — arXiv:2608.18610 — arXiv (en), https://arxiv.org/abs/2608.18610 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e73f8083074a (Metraipolis archive dc_e73f8083074a, sha256 d1afe15a9c96f8080d5c4f4bf8d672b25dda8e76213afe0c5288805d24a2b612)
BibTeX
@misc{dce73f8083074a,
  title = {Denoising-Aware Inversion: Revealing Privacy Risks in Noise-Protected Text Embeddings — arXiv:2608.18610},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.18610},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e73f8083074a; sha256 d1afe15a9c96f8080d5c4f4bf8d672b25dda8e76213afe0c5288805d24a2b612},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.18610. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Dense text embeddings are widely used in data mining, retrieval, and downstream machine learning systems due to their compact and semantically rich representations, but recent embedding inversion attacks have shown that they can expose substantial information about the original text, leading to serious privacy leakage risks. A common defense is to release perturbed embeddings by adding Gaussian noise, which is simple yet effective against standard inversion attacks and does not significantly degrade embedding utility for downstream tasks. However, it remains unclear whether such noise-protected embeddings are sufficiently safe against adaptive attackers that explicitly account for the perturbation process. In this paper, we study text embedding inversion in a noise-protected setting, where the attacker can observe only noisy embeddings and has no access to clean embedding targets. We first analyze why existing generative inversion methods fail under this setting and identify a "Double Noise Trap", which fundamentally prevents standard generative inversion models from achieving high-quality reconstruction. To address this challenge, we propose DAEI, a denoising-aware embedding inversion pipeline that combines a residual denoising autoencoder with generative text inversion where the denoiser is trained in an unsupervised manner using Stein's unbiased risk estimate to enable denoising from noisy observations alone. Extensive experiments show that DAEI achieves approximately 154\% relative improvement in BLEU over the existing generative inversion baseline, while also improving token-level F1 and ROUGE-L by 32--60\%. The promising inversion performance of DAEI challenges the prevailing assumption that simple Gaussian perturbation is sufficient to prevent sensitive information leakage from embedding representations.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-20 in cs.AI).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR).
Pagina del lavoro: https://arxiv.org/abs/2608.18610 — 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 2026-10-18T13:41:17.960Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-20?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).