Online Agnostic Boosting via Regret Minimization
This work solves the problem of boosting in online learning for agnostic settings, which is incremental as it extends existing realizable methods to a broader scenario.
The authors tackled the problem of agnostic online boosting, which previously only existed in realizable settings, by developing the first algorithm that converts a weak learner with marginally-better-than-trivial regret into a strong learner with sublinear regret. They achieved this through a reduction to online convex optimization, unifying statistical and online boosting across agnostic and realizable cases.
Boosting is a widely used machine learning approach based on the idea of aggregating weak learning rules. While in statistical learning numerous boosting methods exist both in the realizable and agnostic settings, in online learning they exist only in the realizable case. In this work we provide the first agnostic online boosting algorithm; that is, given a weak learner with only marginally-better-than-trivial regret guarantees, our algorithm boosts it to a strong learner with sublinear regret. Our algorithm is based on an abstract (and simple) reduction to online convex optimization, which efficiently converts an arbitrary online convex optimizer to an online booster. Moreover, this reduction extends to the statistical as well as the online realizable settings, thus unifying the 4 cases of statistical/online and agnostic/realizable boosting.