Bounding Training Data Reconstruction in Private (Deep) Learning. (arXiv:2201.12383v4 [cs.LG] UPDATED)

Differential privacy is widely accepted as the de facto method for preventing
data leakage in ML, and conventional wisdom suggests that it offers strong
protection against privacy attacks. However, existing semantic guarantees for
DP focus on membership inference, which may overestimate the adversary's
capabilities and is not applicable when membership status itself is
non-sensitive. In this paper, we derive the first semantic guarantees for DP
mechanisms against training data reconstruction attacks under a formal threat
model. We show that two …

data lg training