Private and Online Learnability Are Equivalent-Reference-Cited by-同舟云学术

Private and Online Learnability Are Equivalent

Published:2022-08-16 Issue:4 Volume:69 Page:1-34
ISSN:0004-5411
Container-title:Journal of the ACM
language:en
Short-container-title:J. ACM

Author:

Alon Noga¹,Bun Mark²,Livni Roi³,Malliaris Maryanthe⁴,Moran Shay⁵^ORCID

Affiliation:

1. Princeton University and Tel Aviv University

2. Boston University, Boston, MA, USA

3. Tel Aviv University

4. University of Chicago, IL, USA

5. Technion, Haifa, Israel

Abstract

Let H be a binary-labeled concept class. We prove that H can be PAC learned by an (approximate) differentially private algorithm if and only if it has a finite Littlestone dimension. This implies a qualitative equivalence between online learnability and private PAC learnability.

Funder

NSF

BSF

CAREER

ISF

NSF CAREER

Publisher

Association for Computing Machinery (ACM)

Subject

Artificial Intelligence,Hardware and Architecture,Information Systems,Control and Systems Engineering,Software

Link

https://dl.acm.org/doi/pdf/10.1145/3526074

Reference79 articles.

1. Jacob D. Abernethy, Elad Hazan, and Alexander Rakhlin. 2008. Competing in the dark: An efficient algorithm for bandit linear optimization. In Proceedings of the 21st Annual Conference on Learning Theory (COLT’08). 263–274.

2. Online learning via differential privacy;Abernethy Jacob D.;CoRR,2017

3. Naman Agarwal and Karan Singh. 2017. The price of differential privacy for online learning. In Proceedings of the 34th International Conference on Machine Learning, ICML 2017, Sydney, NSW, Australia, 6–11 August 2017 (Proceedings of Machine Learning Research), Doina Precup and Yee Whye Teh (Eds.). Vol. 70. PMLR, 32–40. http://proceedings.mlr.press/v70/agarwal17a.html.

4. Closure properties for private classification and online prediction;Alon Noga;arXiv preprint arXiv:2003.04509,2020

5. Private PAC learning implies finite Littlestone dimension

Cited by 3 articles. 订阅此论文施引文献订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献

1. No Complete Problem for Constant-Cost Randomized Communication;Proceedings of the 56th Annual ACM Symposium on Theory of Computing;2024-06-10

2. IPOC: An Adaptive Interval Prediction Model based on Online Chasing and Conformal Inference for Large-Scale Systems;Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining;2023-08-04

3. Optimal Differentially Private Learning of Thresholds and Quasi-Concave Optimization;Proceedings of the 55th Annual ACM Symposium on Theory of Computing;2023-06-02