TL;DR
This paper introduces a randomized online paging algorithm that leverages multiple predictors with occasional errors, achieving vanishing regret over time in both full information and bandit models, under the assumption of at least one predictor with sublinear total errors.
Contribution
It is the first to develop an online paging algorithm with vanishing regret using multiple predictors and to address this in both full information and bandit settings.
Findings
The algorithm's regret tends to zero as time increases.
It works effectively even with predictor errors, given at least one predictor with sublinear total errors.
This is the first approach achieving vanishing regret in this context.
Abstract
This paper considers a variant of the online paging problem, where the online algorithm has access to multiple predictors, each producing a sequence of predictions for the page arrival times. The predictors may have occasional prediction errors and it is assumed that at least one of them makes a sublinear number of prediction errors in total. Our main result states that this assumption suffices for the design of a randomized online algorithm whose time-average regret with respect to the optimal offline algorithm tends to zero as the time tends to infinity. This holds (with different regret bounds) for both the full information access model, where in each round, the online algorithm gets the predictions of all predictors, and the bandit access model, where in each round, the online algorithm queries a single predictor. While online algorithms that exploit inaccurate predictions have been…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
Online Paging with a Vanishing Regret· youtube
