Eluder dimension: localise it!
University of Washington · University of Alberta · Google DeepMind / University of Alberta
PDF 由论文原始站点提供，PaperCompass 不保存论文文件。
摘要
We establish a lower bound on the eluder dimension in generalised linear model classes, showing that standard eluder dimension-based analysis cannot lead to first-order regret bounds. To address this, we introduce a localisation method for the eluder dimension; our analysis immediately recovers and improves on classic results for Bernoulli bandits, and allows for the first genuine first-order bounds for finite-horizon reinforcement learning tasks with bounded cumulative returns.