EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Automated Membership Inference via Prompt-Based Attacks in Generative Models

Daniela Gallo, Angelica Liguori, Ettore Ritacco, Luca Caviglione, Fabrizio Durante, Giuseppe Manco

Machine Learning · 2026-05 · 115 (5)

DOI: 10.1007/s10994-026-07010-4 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Machine Learning, online erschienen am 13.05.2026. Verfasst von Daniela Gallo und 5 weiteren Autor:innen. Die Arbeit nennt 50 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Consiglio Nazionale Delle Ricerche.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The growing adoption of prompt-based generative models has raised concerns over the unauthorized use of proprietary data, as such models may memorize and replicate training content. To address this issue, we introduce ProCAP, a novel Membership Inference Attack approach based on a prompt-driven auditing framework.Given a proprietary dataset and a target generative model, ProCAP trains an auxiliary model to craft prompts that trigger the target model to produce outputs revealing potential violations of the proprietary data.Unlike current literature, ProCAP is automatic, fully black-box, model-agnostic, and designed to operate in settings with limited or no knowledge of the training process.To reduce the computational cost of training the prompt generator, we adopt an optimization strategy that filters high-loss samples, i.e., those less likely to have been memorized. Our approach can then “specialize” the learning phase on the most informative data regions. We validate ProCAP across different scenarios, by using both real and synthetic data. Results demonstrate its effectiveness in recognizing unauthorized data usages with strong accuracy-efficiency trade-offs.

Förderung

  • Consiglio Nazionale Delle Ricerche

Zitieren

APA

Gallo, D., Liguori, A., Ritacco, E., Caviglione, L., Durante, F., & Manco, G. (2026). Automated Membership Inference via Prompt-Based Attacks in Generative Models. Machine Learning, 115(5). https://doi.org/10.1007/s10994-026-07010-4

Vancouver

Gallo D, Liguori A, Ritacco E, Caviglione L, Durante F, Manco G. Automated Membership Inference via Prompt-Based Attacks in Generative Models. Machine Learning. 2026;115(5). doi:10.1007/s10994-026-07010-4

BibTeX
@article{Gallo2026,
  author = {Gallo, Daniela and Liguori, Angelica and Ritacco, Ettore and Caviglione, Luca and Durante, Fabrizio and Manco, Giuseppe},
  title = {Automated Membership Inference via Prompt-Based Attacks in Generative Models},
  journal = {Machine Learning},
  year = {2026},
  volume = {115},
  number = {5},
  doi = {10.1007/s10994-026-07010-4}
}
Kostenlos starten →