Logarithmic Regret for Online Gradient Descent Beyond Strong Convexity
This work addresses the challenge of efficient large-scale online learning by providing a memory-efficient algorithm for practitioners, though it is incremental as it builds on classical Hoffman's bound for offline optimization.
The paper tackles the problem of achieving logarithmic regret in online convex optimization over polyhedral sets with curved but not strongly convex loss functions, showing that Online Gradient Descent guarantees logarithmic regret under reasonable data assumptions, with experiments demonstrating comparable or better performance than Online Newton Step.
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 \textit{offline} 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 \textit{online 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 \textit{Online 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 \textit{Online 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 \textit{semi-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.