The MIDST challenge evaluated privacy resilience of diffusion-generated synthetic tabular data via membership inference attacks and produced new black-box and white-box attack methods.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
MIDST Challenge at SaTML 2025: Membership Inference over Diffusion-models-based Synthetic Tabular data
The MIDST challenge evaluated privacy resilience of diffusion-generated synthetic tabular data via membership inference attacks and produced new black-box and white-box attack methods.