English

Anytime Model Selection in Linear Bandits

Machine Learning 2023-11-14 v2 Artificial Intelligence Machine Learning

Abstract

Model selection in the context of bandit optimization is a challenging problem, as it requires balancing exploration and exploitation not only for action selection, but also for model selection. One natural approach is to rely on online learning algorithms that treat different models as experts. Existing methods, however, scale poorly (polyM\text{poly}M) with the number of models MM in terms of their regret. Our key insight is that, for model selection in linear bandits, we can emulate full-information feedback to the online learner with a favorable bias-variance trade-off. This allows us to develop ALEXP, which has an exponentially improved (logM\log M) dependence on MM for its regret. ALEXP has anytime guarantees on its regret, and neither requires knowledge of the horizon nn, nor relies on an initial purely exploratory stage. Our approach utilizes a novel time-uniform analysis of the Lasso, establishing a new connection between online learning and high-dimensional statistics.

Keywords

Cite

@article{arxiv.2307.12897,
  title  = {Anytime Model Selection in Linear Bandits},
  author = {Parnian Kassraie and Nicolas Emmenegger and Andreas Krause and Aldo Pacchiano},
  journal= {arXiv preprint arXiv:2307.12897},
  year   = {2023}
}

Comments

NeurIPS 2023, 37 pages

R2 v1 2026-06-28T11:38:48.132Z