Towards Sharper Utility Bounds for Differentially Private Pairwise Learning
This work addresses privacy concerns in pairwise learning, which is important for applications involving sensitive paired data, though it appears to be an incremental improvement over existing methods.
The paper tackles the problem of providing privacy guarantees for pairwise learning tasks by proposing a new differential privacy paradigm based on gradient perturbation, achieving utility bounds that are similar to or better than previous bounds under convexity assumptions, even for non-convex loss functions.
Pairwise learning focuses on learning tasks with pairwise loss functions, depends on pairs of training instances, and naturally fits for modeling relationships between pairs of samples. In this paper, we focus on the privacy of pairwise learning and propose a new differential privacy paradigm for pairwise learning, based on gradient perturbation. Except for the privacy guarantees, we also analyze the excess population risk and give corresponding bounds under both expectation and high probability conditions. We use the \textit{on-average stability} and the \textit{pairwise locally elastic stability} theories to analyze the expectation bound and the high probability bound, respectively. Moreover, our analyzed utility bounds do not require convex pairwise loss functions, which means that our method is general to both convex and non-convex conditions. Under these circumstances, the utility bounds are similar to (or better than) previous bounds under convexity or strongly convexity assumption, which are attractive results.