Coordinated Online Learning With Applications to Learning User Preferences
Abstract
We study an online multi-task learning setting, in which instances of related tasks arrive sequentially, and are handled by task-specific online learners. We consider an algorithmic framework to model the relationship of these tasks via a set of convex constraints. To exploit this relationship, we design a novel algorithm -- COOL -- for coordinating the individual online learners: Our key idea is to coordinate their parameters via weighted projections onto a convex set. By adjusting the rate and accuracy of the projection, the COOL algorithm allows for a trade-off between the benefit of coordination and the required computation/communication. We derive regret bounds for our approach and analyze how they are influenced by these trade-off factors. We apply our results on the application of learning users' preferences on the Airbnb marketplace with the goal of incentivizing users to explore under-reviewed apartments.
Cite
@article{arxiv.1702.02849,
title = {Coordinated Online Learning With Applications to Learning User Preferences},
author = {Christoph Hirnschall and Adish Singla and Sebastian Tschiatschek and Andreas Krause},
journal= {arXiv preprint arXiv:1702.02849},
year = {2017}
}