Do you pay for Privacy in Online learning?

TL;DR

This paper investigates whether online learning can inherently provide differential privacy without additional mechanisms, exploring the potential equivalence between online learning and privacy guarantees.

Contribution

It formalizes the question of whether privacy can be achieved at no extra cost in online learning, bridging a gap between learning theory and privacy.

Findings

Addresses the theoretical relationship between online learning and differential privacy

Raises the question of whether privacy is achievable without additional effort in online settings

Provides a foundation for future research on privacy-preserving online algorithms

Abstract

Online learning, in the mistake bound model, is one of the most fundamental concepts in learning theory. Differential privacy, instead, is the most widely used statistical concept of privacy in the machine learning community. It is thus clear that defining learning problems that are online differentially privately learnable is of great interest. In this paper, we pose the question on if the two problems are equivalent from a learning perspective, i.e., is privacy for free in the online learning framework?