UMID infers membership in contrastive pre-training data using only text queries by performing latent inversion and comparing similarity and variability signals to synthetic gibberish references via unsupervised anomaly detection.
Membership inference attacks from first principles
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
Generative models for trajectory data do not inherently preserve privacy, as membership inference attacks can identify training data points in representative models.
citing papers explorer
-
Membership Inference for Contrastive Pre-training Models with Text-only PII Queries
UMID infers membership in contrastive pre-training data using only text queries by performing latent inversion and comparing similarity and variability signals to synthetic gibberish references via unsupervised anomaly detection.
-
Privacy Evaluation of Generative Models for Trajectory Generation
Generative models for trajectory data do not inherently preserve privacy, as membership inference attacks can identify training data points in representative models.