Logarithmic Regret for Online Gradient Descent Beyond Strong Convexity

Hoffman's classical result gives a bound on the distance of a point from a convex and compact polytope in terms of the magnitude of violation of the constraints. Recently, several results showed that Hoffman's bound can be used to derive strongly-convex-like rates for first-order methods for textitoffline convex optimization of curved, though not strongly convex, functions, over polyhedral sets. In this work, we use this classical result for the first time to obtain faster rates for textitonline convex optimization over polyhedral sets with curved convex, though not strongly convex, loss functions. We show that under several reasonable assumptions on the data, the standard textitOnline Gradient Descent algorithm guarantees logarithmic regret. To the best of our knowledge, the only previous algorithm to achieve logarithmic regret in the considered settings is the textitOnline Newton Step algorithm which requires quadratic (in the dimension) memory and at least quadratic runtime per iteration, which greatly limits its applicability to large-scale problems. In particular, our results hold for textitsemi-adversarial settings in which the data is a combination of an arbitrary (adversarial) sequence and a stochastic sequence, which might provide reasonable approximation for many real-world sequences, or under a natural assumption that the data is low-rank. We demonstrate via experiments that the regret of OGD is indeed comparable to that of ONS (and even far better) on curved though not strongly-convex losses.