Privately Learning Thresholds: Closing the Exponential Gap

We study the sample complexity of learning threshold functions under the constraint of differential privacy. It is assumed that each labeled example in the training data is the information of one individual and we would like to come up with a generalizing hypothesis $h$ while guaranteeing differential privacy for the individuals. Intuitively, this means that any single labeled example in the training data should not have a significant effect on the choice of the hypothesis. This problem has received much attention recently; unlike the non-private case, where the sample complexity is independent of the domain size and just depends on the desired accuracy and confidence, for private learning the sample complexity must depend on the domain size $X$ (even for approximate differential privacy). Alon et al. (STOC 2019) showed a lower bound of $\Omega(\log^*|X|)$ on the sample complexity and Bun et al. (FOCS 2015) presented an approximate-private learner with sample complexity $\tilde{O}\left(2^{\log^*|X|}\right)$. In this work we reduce this gap significantly, almost settling the sample complexity. We first present a new upper bound (algorithm) of $\tilde{O}\left(\left(\log^*|X|\right)^2\right)$ on the sample complexity and then present an improved version with sample complexity $\tilde{O}\left(\left(\log^*|X|\right)^{1.5}\right)$. Our algorithm is constructed for the related interior point problem, where the goal is to find a point between the largest and smallest input elements. It is based on selecting an input-dependent hash function and using it to embed the database into a domain whose size is reduced logarithmically; this results in a new database, an interior point of which can be used to generate an interior point in the original database in a differentially private manner.

[1]  Kunal Talwar,et al.  Mechanism Design via Differential Privacy , 2007, 48th Annual IEEE Symposium on Foundations of Computer Science (FOCS'07).

[2]  Aaron Roth,et al.  The Algorithmic Foundations of Differential Privacy , 2014, Found. Trends Theor. Comput. Sci..

[3]  Noga Alon,et al.  Private PAC learning implies finite Littlestone dimension , 2018, STOC.

[4]  Guy N. Rothblum,et al.  Boosting and Differential Privacy , 2010, 2010 IEEE 51st Annual Symposium on Foundations of Computer Science.

[5]  Sofya Raskhodnikova,et al.  What Can We Learn Privately? , 2008, 2008 49th Annual IEEE Symposium on Foundations of Computer Science.

[6]  Salil P. Vadhan,et al.  The Complexity of Differential Privacy , 2017, Tutorials on the Foundations of Cryptography.

[7]  Kobbi Nissim,et al.  Differentially Private Release and Learning of Threshold Functions , 2015, 2015 IEEE 56th Annual Symposium on Foundations of Computer Science.

[8]  Amos Beimel,et al.  Private Learning and Sanitization: Pure vs. Approximate Differential Privacy , 2013, APPROX-RANDOM.

[9]  Cynthia Dwork,et al.  Calibrating Noise to Sensitivity in Private Data Analysis , 2006, TCC.

[10]  Vitaly Feldman,et al.  Sample Complexity Bounds on Differentially Private Learning via Communication Complexity , 2014, SIAM J. Comput..

[11]  Moni Naor,et al.  Our Data, Ourselves: Privacy Via Distributed Noise Generation , 2006, EUROCRYPT.

[12]  Thomas Steinke,et al.  Composable and versatile privacy via truncated CDP , 2018, STOC.

[13]  David Haussler,et al.  Occam's Razor , 1987, Inf. Process. Lett..

[14]  Moni Naor,et al.  On the complexity of differentially private data release: efficient algorithms and hardness results , 2009, STOC '09.