Metraipolis › archive › dc_a9ecb032bbf0

Encoded but Exposed: Language Embeddings can Reveal More Than They Should in Recommender Systems — arXiv:2610.21713

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.21713
License (as archived): CC0-1.0
Archived: 2026-10-16T09:28:35.534Z
SHA-256 of the archived text: be46035b30e1780dbcc82cfcb95a167112334861e43148608329089118e4baaf
Stable URL of this page: https://metraipolis.com/documento/dc_a9ecb032bbf0

How to cite

Encoded but Exposed: Language Embeddings can Reveal More Than They Should in Recommender Systems — arXiv:2610.21713 — arXiv (en), https://arxiv.org/abs/2610.21713 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a9ecb032bbf0 (Metraipolis archive dc_a9ecb032bbf0, sha256 be46035b30e1780dbcc82cfcb95a167112334861e43148608329089118e4baaf)

BibTeX

@misc{dca9ecb032bbf0,
  title = {Encoded but Exposed: Language Embeddings can Reveal More Than They Should in Recommender Systems — arXiv:2610.21713},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.21713},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a9ecb032bbf0; sha256 be46035b30e1780dbcc82cfcb95a167112334861e43148608329089118e4baaf},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.21713. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Language embeddings are increasingly used in recommenders to represent item content, but their privacy risks remain poorly understood. We study whether item embeddings can reveal item identities when an adversary has access to auxiliary item text from external sources. We show that a simple cosine-similarity attack can recover item identities with high mean reciprocal rank across language models and candidate-set settings on two real-world datasets. Leakage is strongest when the adversary's partial information overlaps with the victim representation, but it also persists through semantic alignment between independently sourced descriptions. We further show that item leakage can propagate to user re-identification from sparse ratings. Finally, we evaluate obfuscated descriptions as a mitigation and find a clear leakage-utility tradeoff.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Cryptography and Security (cs.CR) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.21713 — 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 2026-10-16T08:44:17.986Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).