BibTeX
CSL-JSON
MLA
Harvard
An Equivalence Between Private Classification and Online Prediction release_xxr6uppdevfk7e7pesxlvglkje
by Mark Bun and Roi Livni and Shay Moran
Released as a article .
2021
Abstract
We prove that every concept class with finite Littlestone dimension can be learned by an (approximate) differentially-private algorithm. This answers an open question of Alon et al. (STOC 2019) who proved the converse statement (this question was also asked by Neel et al.~(FOCS 2019)). Together these two results yield an equivalence between online learnability and private PAC learnability. We introduce a new notion of algorithmic stability called "global stability" which is essential to our proof and may be of independent interest. We also discuss an application of our results to boosting the privacy and accuracy parameters of differentially-private learners.
In text/plain
format
Archived Files and Locations
application/pdf 589.1 kB file_lswk6hbnb5bojbvsqendp6b7ji |
arxiv.org (repository) web.archive.org (webarchive) |
Read Archived PDF
Preserved and Accessible
arXiv
2003.00563v3
Work Entity
access all versions, variants, and formats of this works (eg, pre-prints)
access all versions, variants, and formats of this works (eg, pre-prints)
Cite This