Timezone: »

Differential Privacy without Sensitivity
Kentaro Minami · Hiromi Arai · Issei Sato · Hiroshi Nakagawa

Wed Dec 07 09:00 AM -- 12:30 PM (PST) @ Area 5+6+7+8 #6
The exponential mechanism is a general method to construct a randomized estimator that satisfies $(\varepsilon, 0)$-differential privacy. Recently, Wang et al. showed that the Gibbs posterior, which is a data-dependent probability distribution that contains the Bayesian posterior, is essentially equivalent to the exponential mechanism under certain boundedness conditions on the loss function. While the exponential mechanism provides a way to build an $(\varepsilon, 0)$-differential private algorithm, it requires boundedness of the loss function, which is quite stringent for some learning problems. In this paper, we focus on $(\varepsilon, \delta)$-differential privacy of Gibbs posteriors with convex and Lipschitz loss functions. Our result extends the classical exponential mechanism, allowing the loss functions to have an unbounded sensitivity.

Author Information

Kentaro Minami (The University of Tokyo)
Hiromi Arai (The University of Tokyo)
Issei Sato (The University of Tokyo)
Hiroshi Nakagawa (The University of Tokyo)

More from the Same Authors