Adaptive Online Learning in Dynamic Environments
This work addresses the challenge of adaptive online learning for dynamic environments, providing an optimal solution for scenarios where comparators change over time, which is incremental but improves upon existing bounds.
The paper tackles the problem of online convex optimization in dynamic environments by developing a novel method called Ader, which achieves an optimal O(√(T(1+P_T))) dynamic regret, closing the gap from the previous O(√(T)(1+P_T)) bound to match the established lower bound.
In this paper, we study online convex optimization in dynamic environments, and aim to bound the dynamic regret with respect to any sequence of comparators. Existing work have shown that online gradient descent enjoys an $O(\sqrt{T}(1+P_T))$ dynamic regret, where $T$ is the number of iterations and $P_T$ is the path-length of the comparator sequence. However, this result is unsatisfactory, as there exists a large gap from the $Ω(\sqrt{T(1+P_T)})$ lower bound established in our paper. To address this limitation, we develop a novel online method, namely adaptive learning for dynamic environment (Ader), which achieves an optimal $O(\sqrt{T(1+P_T)})$ dynamic regret. The basic idea is to maintain a set of experts, each attaining an optimal dynamic regret for a specific path-length, and combines them with an expert-tracking algorithm. Furthermore, we propose an improved Ader based on the surrogate loss, and in this way the number of gradient evaluations per round is reduced from $O(\log T)$ to $1$. Finally, we extend Ader to the setting that a sequence of dynamical models is available to characterize the comparators.