Simple Stochastic and Online Gradient Descent Algorithms for Pairwise Learning

Abstract

Pairwise learning refers to learning tasks where the loss function depends on a pair of instances. It instantiates many important machine learning tasks such as bipartite ranking and metric learning. A popular approach to handle streaming data in pairwise learning is an online gradient descent (OGD) algorithm, where one needs to pair the current instance with a buffering set of previous instances with a sufficiently large size and therefore suffers from a scalability issue. In this paper, we propose simple stochastic and online gradient descent methods for pairwise learning. A notable difference from the existing studies is that we only pair the current instance with the previous one in building a gradient direction, which is efficient in both the storage and computational complexity. We develop novel stability results, optimization, and generalization error bounds for both convex and nonconvex as well as both smooth and nonsmooth problems. We introduce novel techniques to decouple the dependency of models and the previous instance in both the optimization and generalization analysis. Our study resolves an open question on developing meaningful generalization bounds for OGD using a buffering set with a very small fixed size. We also extend our algorithms and stability analysis to develop differentially private SGD algorithms for pairwise learning which significantly improves the existing results.

Error analysis of kernel regularized pairwise learning with a strongly convex loss

“…[25]) et al During the past decades, various methods for pairwise learning have been proposed, the performance of the kernel-based regularized pairwise learning models has been extensively studied (see e.g. [20,24,33,34,46,48,54]). For instance, [15] researches the learning rates of a pairwise ranking model stated in the framework of misranking loss and empirical risk minimization, [37] extends the result to general convex loss functions, and later, [59] improves the learning rates for regularized least squares ranking algorithm.…”